| Message ID | 1472838001-1673-1-git-send-email-peter.maydell@linaro.org |
|---|---|
| State | Accepted |
| Commit | 17351c3f11e52101e1498ba03eec06fa5e4e713a |
| Headers | show |
Ping ? thanks -- PMM On 2 September 2016 at 18:40, Peter Maydell <peter.maydell@linaro.org> wrote: > The sigevent structure includes a union with some fields which > are pointers. For the QEMU target_sigevent structure we must > represent these as abi_ulongs, not host function pointers. > > This error was causing the compiler to believe it should 8-align > the _sigev_un union on a 64-bit host, which meant that the > code in target_to_host_sigevent() was looking at the wrong > offset to find the _tid field, and timer_create() would > spuriously fail with EINVAL. > > This fixes the final loose end noted in LP:1042388. > > While we're editing the structure, switch the 'int32_t' fields > to 'abi_int'; this will only matter for guests with non-standard > integer alignment like m68k. > > Signed-off-by: Peter Maydell <peter.maydell@linaro.org> > --- > linux-user/syscall_defs.h | 16 ++++++++++------ > 1 file changed, 10 insertions(+), 6 deletions(-) > > diff --git a/linux-user/syscall_defs.h b/linux-user/syscall_defs.h > index dbf6a38..f4417ee 100644 > --- a/linux-user/syscall_defs.h > +++ b/linux-user/syscall_defs.h > @@ -2618,15 +2618,19 @@ typedef int32_t target_timer_t; > > struct target_sigevent { > target_sigval_t sigev_value; > - int32_t sigev_signo; > - int32_t sigev_notify; > + abi_int sigev_signo; > + abi_int sigev_notify; > union { > - int32_t _pad[TARGET_SIGEV_PAD_SIZE]; > - int32_t _tid; > + abi_int _pad[TARGET_SIGEV_PAD_SIZE]; > + abi_int _tid; > > + /* The kernel (and thus QEMU) never looks at these; > + * they're only used as part of the ABI between a > + * userspace program and libc. > + */ > struct { > - void (*_function)(sigval_t); > - void *_attribute; > + abi_ulong _function; > + abi_ulong _attribute; > } _sigev_thread; > } _sigev_un; > }; > -- > 2.7.4
On Tue, Oct 04, 2016 at 02:08:24PM +0100, Peter Maydell wrote: > Ping ? Applied, thanks. > thanks > -- PMM > > On 2 September 2016 at 18:40, Peter Maydell <peter.maydell@linaro.org> wrote: > > The sigevent structure includes a union with some fields which > > are pointers. For the QEMU target_sigevent structure we must > > represent these as abi_ulongs, not host function pointers. > > > > This error was causing the compiler to believe it should 8-align > > the _sigev_un union on a 64-bit host, which meant that the > > code in target_to_host_sigevent() was looking at the wrong > > offset to find the _tid field, and timer_create() would > > spuriously fail with EINVAL. > > > > This fixes the final loose end noted in LP:1042388. > > > > While we're editing the structure, switch the 'int32_t' fields > > to 'abi_int'; this will only matter for guests with non-standard > > integer alignment like m68k. > > > > Signed-off-by: Peter Maydell <peter.maydell@linaro.org> > > --- > > linux-user/syscall_defs.h | 16 ++++++++++------ > > 1 file changed, 10 insertions(+), 6 deletions(-) > > > > diff --git a/linux-user/syscall_defs.h b/linux-user/syscall_defs.h > > index dbf6a38..f4417ee 100644 > > --- a/linux-user/syscall_defs.h > > +++ b/linux-user/syscall_defs.h > > @@ -2618,15 +2618,19 @@ typedef int32_t target_timer_t; > > > > struct target_sigevent { > > target_sigval_t sigev_value; > > - int32_t sigev_signo; > > - int32_t sigev_notify; > > + abi_int sigev_signo; > > + abi_int sigev_notify; > > union { > > - int32_t _pad[TARGET_SIGEV_PAD_SIZE]; > > - int32_t _tid; > > + abi_int _pad[TARGET_SIGEV_PAD_SIZE]; > > + abi_int _tid; > > > > + /* The kernel (and thus QEMU) never looks at these; > > + * they're only used as part of the ABI between a > > + * userspace program and libc. > > + */ > > struct { > > - void (*_function)(sigval_t); > > - void *_attribute; > > + abi_ulong _function; > > + abi_ulong _attribute; > > } _sigev_thread; > > } _sigev_un; > > }; > > -- > > 2.7.4 >
diff --git a/linux-user/syscall_defs.h b/linux-user/syscall_defs.h index dbf6a38..f4417ee 100644 --- a/linux-user/syscall_defs.h +++ b/linux-user/syscall_defs.h @@ -2618,15 +2618,19 @@ typedef int32_t target_timer_t; struct target_sigevent { target_sigval_t sigev_value; - int32_t sigev_signo; - int32_t sigev_notify; + abi_int sigev_signo; + abi_int sigev_notify; union { - int32_t _pad[TARGET_SIGEV_PAD_SIZE]; - int32_t _tid; + abi_int _pad[TARGET_SIGEV_PAD_SIZE]; + abi_int _tid; + /* The kernel (and thus QEMU) never looks at these; + * they're only used as part of the ABI between a + * userspace program and libc. + */ struct { - void (*_function)(sigval_t); - void *_attribute; + abi_ulong _function; + abi_ulong _attribute; } _sigev_thread; } _sigev_un; };
The sigevent structure includes a union with some fields which are pointers. For the QEMU target_sigevent structure we must represent these as abi_ulongs, not host function pointers. This error was causing the compiler to believe it should 8-align the _sigev_un union on a 64-bit host, which meant that the code in target_to_host_sigevent() was looking at the wrong offset to find the _tid field, and timer_create() would spuriously fail with EINVAL. This fixes the final loose end noted in LP:1042388. While we're editing the structure, switch the 'int32_t' fields to 'abi_int'; this will only matter for guests with non-standard integer alignment like m68k. Signed-off-by: Peter Maydell <peter.maydell@linaro.org> --- linux-user/syscall_defs.h | 16 ++++++++++------ 1 file changed, 10 insertions(+), 6 deletions(-) -- 2.7.4